Frame by frame transfer, keeping it simple

[Tony Panciroli]

hi. i introduced myself the other day so thought i would get stuck in and fire up a post of what i am doing, i am an IT engineer but i wanted to find the simplest way to do a frame by frame capture of super 8 and standard 8. i got a Raspberry pi and HQ camera along with microscope macro c Mount lens and then sat it all in place on a projector already converted by moviestuff. so it has an LED lamp and speed controller on the motor.
I programmed the Pi to take a picture on button press and then added in a limit switch and a cam onto the shutter wheel of the projector. so that every revolution of the wheel it takes a picture. and timed it so that it is right after the projector moves onto the next frame.
now this projector is not ideal. there is not enough room to get square onto the gate so i knew this would be my test bench machine but it worked. i have uploaded a bit of film that i have done as a test. only thing done is cropped and flipped.
Also at this point i have not removed the projectors shutters. so there is flicker.. just ignore that as a test

as it was pictured. couldnt get closer with lens because of the design of the projector
https://youtu.be/7BKbFa6o7NA

after flip and crop no other processes carried out
https://youtu.be/XEpOzy6Qctw

https://youtu.be/90Sgfv_yX6Q

https://youtu.be/7cbI8B4eCwM


I have now just gotten a Sankyo Dualux-8 Super 8 Film Projector.
This looks so far much better suited to what i need.

https://photos.app.goo.gl/AhCU453sJK5s24i9A

I will keep you posted how it unfolds

[Tony Panciroli]

i have now removed the old electric motor and transformer.
At low speed it did not have the power to keep to a consistant speed.
So ive removed it in favor of a stepper motor and toothed belt drive.
Just waiting on the driver and controller to test.
i have also cut off the shutters from the projector and turned the still frame clutch system into a mechanism that engages the limit switch to operate the camera shutter. it is timed so that it takes a picture at the end of the claws stroke. which gives the camera time to process before changing frame again.
Pictures show before and after motor conversion

[Tony Panciroli]

So thats the machine finished. and i am delighted with the results. No complec programming and timing issues. most of the set up is mechanical
i am delighted with the results.
only thing i have changed is i swapped out the limit switch for a magnetic sensor. which triggers every revolution just after the claw moves the film. the result is rock steady. colours are good and ive put through 1500 foot so far and its not missed a beat. total price around £100 GBP.

https://youtu.be/JH70rlSspF8


https://youtu.be/nssMrG1_C_8

I have a couple of parts to print to tidy it all up now. and may put a dimmer on the light as its maybe a tiny bit too strong.

[Janice Glesser]

Great job Tony. The results are very impressive.

[Tony Panciroli]

thank you. i was determined to build one. but it seams that so many are over complicating it. having to code this that and the next. i am in IT and i know how to code. but sometimes things just need a mechanical approach. couple that with some mechanno and its all good.

[Stan Jelavic]

Nice work Tony. My only note is that the resolution seems to be on the low side.
See my clip done with the Wolverine/Hawkeye unit in comparison.
https://drive.google.com/file/d/1ytl...usp=share_link
This uses the imaging source camera with the 12mm good quality lens.
Another note, if you want to go higher scan rates and without flicker then the things get more complex. I see you already tried that.

[Andrew Davis]

Hi Tony, thank you for writing about this. I'm interested in doing something similar. Could you give more details of the microscope lens that you are using - the specifications, where you bought it, etc? Also, from the photo, it looks like you have the Pi camera mounted on a stand - is that free-standing? I'm guessing the focussing is tricky. Does the lens fill the sensor in the camera? I've tried various camera/lens combinations but never with great results so I would be grateful to know more about your optical set-up.

[Tony Panciroli]

Nice work Tony. My only note is that the resolution seems to be on the low side.
See my clip done with the Wolverine/Hawkeye unit in comparison.
https://drive.google.com/file/d/1ytl...usp=share_link
This uses the imaging source camera with the 12mm good quality lens.
Another note, if you want to go higher scan rates and without flicker then the things get more complex. I see you already tried that.

it will scan the frames at whatever resolution i set it to. up to 2k. but as the film is not sharp to begin with so i kept it lower.
Also remember that Colour film was just released when this was taken. your film although very good is a LOT newer. by the looks about 20 years newer.

i dont need it to scan quickly i have all the time i need. i have upped the rate to 2 frames per second and that is plenty.
The image with flicker was before cutting the fins off in the projector. i have no flicker now

these are now coming out way better than the results we got for paying to have them done. so i am more than happy with the results now.

[Stan Jelavic]

Was curious if the reason for lower rez was the lens or the film. Have you tried any more recent good quality film? My film is from late 70's.

[Richard Stanton]

Love how stable that is. My (bought not built) "movie maker pro" converter jumps about so much it needs anti-shake software for post processing

[Bill Cain]

Finally a rational simple design. Great job!
The only thing that worries me with my limited ability to program is the Raspberry PI. It is very intimidating actually. It seems they are unavailable at the time. The few who have them are price gouging currently. I am leaning toward the mirrorless camera/Laowa macro lens mechanically triggered method. Its also hard to find a used DSLR at a reasonable price.
Again, nice work.

[Tony Panciroli]

Finally a rational simple design. Great job!
The only thing that worries me with my limited ability to program is the Raspberry PI. It is very intimidating actually. It seems they are unavailable at the time. The few who have them are price gouging currently. I am leaning toward the mirrorless camera/Laowa macro lens mechanically triggered method. Its also hard to find a used DSLR at a reasonable price.
Again, nice work.

Thank you. it is what i was going for. i have been following a lot of other people across the world in their efforts. and although some very impressive results are out there they are very complex. and some of the programming and synchronisation is just so complex.
I have been doing cine since i was a child with my father. and air more on the mechanical side over the programming side. and i just broke it down to every frame i want it to take a picture. SLR is great and would have been right on my list but the mirror will not last being triggered so many times.
so i had a pi sitting around. and just went from there. just very simple. bearly any programming. and its triggered by a magnetic sensor.

[Stan Jelavic]

I like your approach Tony to make it simple but do have a few additional questions. Looks like the lens that you used is the one that was used with the t-scan project.
https://tscann8.torulf.com/v2/index.html
https://shop.pimoroni.com/products/m...31885093371987
My original concern was resolution. From the t-scan sample videos I can se that the resolution seems to be pretty good but I see chromatic aberrations present. Not very noticeable but it is there.
Torulf discussed this here:
https://tscann8.torulf.com/v2/hq-update.html
He is using daVinci Resolve to reduce it and seems pretty happy with the solution.
I was thinking of going this way with one of my projects but wanted to make sure I will be getting good results and that is why I am asking these questions. On the other hand it is not a big money and I may just go and try it.

[Tony Panciroli]

Hi. sorry for very late response.
i feel the resolution is fantastic. and i am delighted with my results.
and in fact when i watch them on my 65inch OLED and project the same size my results look better than what comes out the projector.
i feel thats all i need. i do not need to be more critical than that.
i do not enhance them in any way either or remove disc etc for the very reason you would listen to Records over a CD. you are not looking for clarity. its a whole atmosphere. you already detract from it in removing the projector.
my advice is just go with what you feel is a good direction. and if you are happy with your results then you have accomplished your goal.
you dont need others picking it apart for you.

[Tony Panciroli]

THe t-scan project is great. very impressive. i do a lot of 3D printing as well. but as with any project i take on i sit back and ask myself if it exists already.
In this case he has built the whole scanner bed from scratch basically to mimic what a projector is doing......... well a projector already does this. and moves the film frame by frame. so my mind goes to how can i convert something to work the way i want it to. my project took me 3 weeks to build. and in fact i could have done it in a day if i didn't have to wait for parts. impressive as the from scratch builds are they dont do anything that mine does not. i dont even require code to centre the frame. as the mechanical claw does it perfect on every stroke